Latency-associated Nuclear Antigen of Kaposi's Sarcoma-associated Herpesvirus Functionally Interacts with Heterochromatin Protein 1*

Latency-associated nuclear antigen (LANA) of Kaposi's sarcoma-associated herpesvirus plays an important role in maintenance of the viral genome during latent infection. LANA additionally participates in the transcriptional regulation of several viral and cellular promoters. When tethered to constitutively active promoters, the protein exhibits transcriptional repressor activity. In this report, we further characterized cell type-, promoter-, and domain-specific transcriptional repression by LANA. We additionally speculated on the mechanism underlying transcriptional repression by the C terminus of the protein. Subnuclear localization patterns and association with heterochromatin suggested a possible link between LANA and heterochromatin protein 1, a representative heterochromatin-associated protein. In vivo and in vitro binding and immunofluorescence assays revealed that LANA associates with heterochromatin protein 1 in an isotype-specific manner. Furthermore, biochemical fractionation and transient replication assays supported the possibility that this interaction contributes to transcriptional repression, targeting to subnuclear structures, and latent DNA replication activity of LANA.

Latency-associated nuclear antigen (LANA) of Kaposi's sarcoma-associated herpesvirus plays an important role in maintenance of the viral genome during latent infection. LANA additionally participates in the transcriptional regulation of several viral and cellular promoters. When tethered to constitutively active promoters, the protein exhibits transcriptional repressor activity. In this report, we further characterized cell type-, promoter-, and domain-specific transcriptional repression by LANA. We additionally speculated on the mechanism underlying transcriptional repression by the C terminus of the protein. Subnuclear localization patterns and association with heterochromatin suggested a possible link between LANA and heterochromatin protein 1, a representative heterochromatin-associated protein. In vivo and in vitro binding and immunofluorescence assays revealed that LANA associates with heterochromatin protein 1 in an isotypespecific manner. Furthermore, biochemical fractionation and transient replication assays supported the possibility that this interaction contributes to transcriptional repression, targeting to subnuclear structures, and latent DNA replication activity of LANA.
The nonhistone chromosomal protein, heterochromatin protein 1 (HP1), 1 is tightly associated with heterochromatin. It was originally identified by expression library screening with a monoclonal antibody against a fraction of DNA-binding nuclear proteins of Drosophila melanogaster (1). The chromosomal rearrangement of euchromatic genes adjacent to the heterochromatic regions results in gene silencing, designated "position effect variegation." Mutations in a Drosophila suppressor of the variegation 2-5 gene encoding HP1 result in dosage-dependent dominant suppression of position effect variegation (2,3). In addition, HP1 is an essential gene in Drosophila (3). Its mutation affects the segregation of chromosomes (4) and causes multiple telomeric fusions (5). HP1 homologs have been iden-tified in diverse organisms, from fission yeast to mammals (for reviews see Refs. 6 and 7). Three isotypes of mammalian HP1 protein, denoted HP1␣, HP1␤, and HP1␥, have been reported (8 -11). All isotypes contain common structural motifs, specifically an N-terminal chromo domain and C-terminal chromo shadow domain connected by a hinge region. However, their subnuclear localization and phosphorylation states during the cell cycle are distinct (12,13), suggesting participation in different chromosomal events. Analysis of a conserved sequence motif in Drosophila Polycomb protein and HP1 led to the identification of a chromo domain termed for chromosome organization modifier (14), and a second chromo domain-like motif in HP1, known as chromo shadow domain (15) that mediates most protein-protein interactions, including self-association (6, 7, 10, 16 -18). The chromo domain is present in diverse proteins that are potentially involved in chromosomal organization (18 -20). Recent reports (21,22) suggest that the chromo domain of HP1 is responsible for the recognition of methylated lysine 9 on histone H3. These results, in addition to the finding that HP1 physically interacts with suv39h1, a histone methyltransferase responsible for the methylation of lysine 9 on histone H3 (23,24), suggests a self-propagating mechanism of heterochromatin. Interestingly, suv39h1-HP1 complexes are involved not only in heterochromatic gene silencing but also in the transcriptional regulation of the euchromatic cyclin E promoter (25,26), which was shown previously to be regulated partly by the recruitment of histone deacetylase via a retinoblastoma protein (27). These data collectively implicate an additional molecular mechanism supporting the histone code hypothesis that differential and combinatorial post-translational modifications of the N-terminal tail of histone, including acetylation, methylation, and phosphorylation, epigenetically mark the chromosomal state as open or closed for transcriptional activity (28). However, the enzymes responsible for histone demethylation or removal of methylated histone tail remain to be identified (29).
Kaposi's sarcoma is the most common neoplasm in persons with acquired immunodeficiency syndrome. Kaposi's sarcomaassociated herpesvirus (KSHV) was originally identified from Kaposi's sarcoma tissues of acquired immunodeficiency syndrome patients, using representational difference analyses (30), and later shown to associate with several lymphoproliferative diseases, including body cavity-based lymphoma/primary effusion lymphoma and some cases of multicentric Castleman's disease (31,32). KSHV switches its infection cycle between latent and lytic states. During latent infection, viral gene expression is restricted to a small subset of the entire open reading frame of the KSHV genome (33), and the circularized viral genome is maintained as a multicopy episome (34,35) by the latency-associated nuclear antigen (LANA) encoded by open reading frame 73 (36,37). LANA associates with the mitotic chromosome (38 -40) and binds to sites within the 801-bp terminal repeat (TR) sequences of the viral genome (41)(42)(43)(44)(45), thereby allowing the maintenance of plasmids containing this region in the long term selection of drug resistance of stably transfected cells (38,41). Recently, we and others (44,45) reported that LANA is required for the DNA replication of KSHV TR-containing plasmids, based on data from a transient replication assay using a methylation-sensitive restriction enzyme. Moreover, the protein interacts with components of origin recognition complexes (ORCs) (45), similar to Epstein-Barr virus nuclear antigen-1, a functional analog of KSHV LANA in the latent replication of the viral genome (46 -48). These data suggest that LANA maintains the KSHV genome, not only by tethering the viral episome to host chromosomes during host cell mitosis but also by actively participating in the DNA replication of the viral genome, possibly through interactions with cellular replication machinery, such as ORCs. In addition to maintenance of the viral genome, LANA interacts with several cellular transcription factors (49 -53) and regulates their activities in cellular and viral promoters (43, 50, 52, 54 -58). LANA represses transcription from KSHV TR by direct binding (43)(44)(45), and exerts transcriptional repression activity when tethered to a heterologous promoter via the GAL4 DNA-binding domain (DBD) (50,59). We are interested in determining the molecular mechanism underlying the transcriptional repression activity of LANA. In this report, we speculate on the functional association between LANA and HP1, and the consequent effects on transcriptional activity, subnuclear structure targeting, and latent replication activity.

EXPERIMENTAL PROCEDURES
Plasmids-Plasmids pcDNA3 LANA, pFLAG-CMV2 LANA, pEBG LANA, and their derivatives encoding deletion mutants of LANA were described previously (45,51,53). Samples of cDNA corresponding to specific regions of LANA were amplified from pcDNA3 LANA by PCR, with appropriate sets of primers, and inserted into CMV G4 (45) to express GAL4 DBD-fused proteins in mammalian cells. The pGEX2T hHP1␣ plasmid (60) was a gift from Dr. A. Dejean (Institut Pasteur, France). pGEX2T mHP1␣, mHP1␤, and mHP1␥ (61) were generously supplied by Dr. R. Losson (Institut de Genetique et de Biologie Moleculaire et Cellulaire, France). A BamHI/EcoRI fragment of pGEX2T hHP1␣ was inserted into the corresponding restriction sites of pcDNA3 (Invitrogen) and the BglII/EcoRI sites of pEGFP C1 (Clontech, Palo Alto, CA). A BamHI/NotI fragment of pcDNA3 hHP1␣ was inserted in the corresponding sites of pEBG to express glutathione S-transferase (GST)-fused hHP1␣ in mammalian cells. cDNAs corresponding to hHP1␣, mHP1␤, and mHP1␥ were amplified by PCR with appropriate sets of primers and inserted into the EcoRI/XhoI site of pME18S (62) to express FLAG-tagged proteins in mammalian cells. Deletion mutants of hHP1␣ were constructed using similar procedures. pFR-luc was purchased from Stratagene (San Diego, CA). A BamHI fragment of pFR-luc containing five tandem GAL4-DBD-binding sites was inserted into the BglII site of the pGL3 promoter (Promega, Madison, WI), and the resulting construct was designated pGx5SV-luc. pG4M2-luc was a gift from Dr. H. G. Stunnenberg (University of Nijmegen, The Netherlands). The p4TR-luc construct was described earlier (45).
Cell Culture, Transfection, and Reporter Assay-293T cells were maintained and transfected as described previously (51). The human cervical cancer cell line, C33A, was maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and transfected using the calcium phosphate method. The quantity of total DNA used in transfection was kept constant by including an appropriate blank vector. The transient reporter assay was performed as described in a previous report (51).
In Vivo and in Vitro Binding Assays-293T cells in 100-mm dishes were cotransfected with specific combinations of expression vectors. After ϳ36 h of transfection, cells were harvested and pellets stored at Ϫ70°C before use. Cells were lysed in 500 l of ice-cold phosphatebuffered saline containing 0.5% Nonidet P-40 and 1 mM phenylmethylsulfonyl fluoride, with brief sonication. Cell debris was removed by centrifugation. For coimmunoprecipitation, the supernatant was incubated with anti-FLAG M2 monoclonal antibody (Sigma) at 4°C for 1 h. Following the addition of protein G-Sepharose (Amersham Biosciences), the supernatant was further incubated at 4°C for 3 h. For the GST pull-down assay, the supernatant was incubated with glutathione-Sepharose 4B (Amersham Biosciences) at 4°C for 3 h. Beads were washed three times in 1 ml of lysis buffer, and bound proteins were eluted with SDS gel loading buffer. Eluted proteins were separated by SDS-PAGE, transferred to a nitrocellulose membrane, immunoblotted with the indicated antibodies, and detected using ECL TM (Amersham Biosciences). The in vitro binding assay was performed as described previously (45).
Immunofluorescence Microscopy-293T cells grown on coverslips were cotransfected with the specified combinations of expression vectors. At ϳ24 h post-transfection, cells were either fixed in methanol at Ϫ20°C for 10 min or 3.7% formaldehyde at room temperature for 30 min and permeabilized in phosphate-buffered saline containing 0.2% Triton X-100 at 4°C for 25 min. LANA was detected with rabbit polyclonal anti-LANA serum (a gift from Dr. J. Jung, Harvard Medical School) and fluorescein isothiocyanate-conjugated goat anti-rabbit antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). FLAG-tagged proteins were observed using an anti-FLAG M2 monoclonal antibody (Sigma) and rhodamine-conjugated goat anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories). Coverslips were mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA) and examined by confocal laser scanning microscopy (Pascal, Carl Zeiss, Jena, Germany).
Fractionation of Cellular Proteins-Subcellular fractionation was performed as described previously (63) with minor modifications. 293T cells in 60-mm dishes were transfected with expression vectors containing GAL4 DBD fusion proteins and harvested ϳ36 h after transfection. Cells were lysed in cytoskeleton buffer (10 mM PIPES (pH 6.8), 100 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 1 mM EGTA, 1 mM dithiothreitol, 0.1 mM ATP, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , and 10 mM NaF) containing 0.1% Triton X-100 at 4°C for 15 min. After centrifugation at 4000 rpm for 3 min, soluble fractions extracted by Triton X-100 were further clarified by centrifugation at 12000 rpm for 15 min. Triton-extracted nuclei were washed once in cytoskeleton buffer containing 0.1% Triton X-100 and subjected to chromatin digestion in the above buffer containing 1000 units/ml DNase I (Sigma) at room temperature for 30 min. DNase I-extractable and -resistant proteins were separated by low speed centrifugation. A schematic representation of subcellular fractionation is depicted in Fig. 5A. GAL4 DBD fusion proteins were detected with anti-GAL4 DBD monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA).
Transient Replication Assay-The transient replication assay was performed as described previously (45). C33A cells on 100-mm dishes were cotransfected with 2 g of p4TR-luc, 2 g of pGL2-basic as an internal control, and 8 g of FLAG-CMV2 derivatives expressing FLAG-tagged LANA mutants. At ϳ36 h post-transfection, cells were trypsinized, collected by centrifugation, and diluted into fresh media. Cells were harvested 96 h after transfection, and low molecular weight DNA was extracted by the Hirt lysis method (64). Supernatants were successively extracted with phenol/chloroform/isoamyl alcohol and chloroform. Ethanol-precipitated DNA was dissolved in 50 l of distilled water containing RNase. DpnI and Alw44I were used to digest unreplicated DNA and linearize pGL2-basic derivatives, respectively. DpnI digestion was confirmed by monitoring the complete digestion of the unreplicated internal control, pGL2-basic. Southern blot hybridization and detection of plasmids containing the luciferase gene were performed as described previously (45).

Promoter-and Cell Type-specific Transcriptional Repression
Domain of KSHV LANA-Earlier reports (50,59) suggest that LANA functions as a transcriptional repressor when tethered to constitutively active promoters via a heterologous DNAbinding domain. However, the domain of LANA responsible for transcriptional repression remains to be conclusively identified. Schwam et al. (59) demonstrated that both the N and C termini of LANA contain a transcriptional repression domain, in contrast to reports by Krithivas et al. (50) that only the N terminus of LANA mediates transcriptional repression, possibly via interactions with the mSin3 corepressor complex. Because different cell lines and reporters were employed in the two sets of experiments, it is possible that the cellular or promoter context affected the transcriptional activity of LANA. Accordingly, we examined the transcriptional activity of fulllength LANA and derivatives fused to GAL4 DBD in a transient reporter assay, using two cell lines and four different reporters containing tandem repeats of GAL4 DBD-binding sites. The reporters include pFR-luc, pG4M2-luc, pGx5SV-luc, and pGal4TK-luc containing the basic promoter element (TATA box), adenovirus major late promoter, SV40 promoter, and thymidine kinase promoter downstream of the tandem GAL4 DBD-binding sites, respectively. As shown in Fig. 1A, both full-length and N-terminal LANA corresponding to amino acids 1-340 fused to GAL4 DBD displayed comparable transcriptional repressor activity on all reporters tested in transiently transfected 293T cells. However, transcriptional repression by C-terminal LANA corresponding to amino acids 950 -1162 fused to GAL4 DBD was promoter-specific; pG4M2-luc was slightly repressed, and pGX5SV-luc was not affected. LANA and deletion mutants not fused to GAL4 DBD included in the same reporter assay did not affect the transcriptional activity of reporters (data not shown). Therefore, it seems unlikely that the observed promoter specificity is because of the effect of C-terminal LANA itself on transcription from each promoter independently of GAL4 DBD sites. Similar transient reporter assays were performed in C33A cells (Fig. 1B). Unexpectedly, the N terminus of LANA fused to GAL4 DBD did not display transcriptional repression activity on any of the report-ers tested. In contrast, the C terminus of LANA fused to GAL4 DBD repressed transcriptional activity with similar promoter specificity as that observed in 293T cells. To determine the transcriptional repression domain within the C terminus of LANA, we constructed a series of deletion mutants fused to GAL4 DBD (Fig. 1C). As shown in Fig. 1D, the central region of the C terminus is responsible for transcriptional repression in both cell lines. The inability of LANA C⌬161 and LANA C1107 fused to GAL4 DBD to repress transcription was not due to lower expression levels in transfected cells, as verified by Western blotting with anti-GAL4 DBD antibody (data not shown). The results collectively suggest that both the N and C termini of LANA display cell type-and promoter-specific transcriptional repression activity.
In Vivo and in Vitro Binding of KSHV LANA to HP1-KSHV LANA associates preferentially with the border of heterochromatin in interphase nuclei (65). Because HP1 is a representative nonhistone chromosomal protein tightly associated with heterochromatin, we speculate that the association of LANA with heterochromatin is mediated by interactions with HP1. Additionally, the transcriptional repression activity of GAL4 DBD-fused LANA may be partly explained by a gene silencing DED, aspartate/glutamate-rich repeat region; Q-rich, glutamine-rich region; ZIP, putative leucine-zipper domain. D, 293T cells in 60-mm dishes were cotransfected with 1 g of indicated reporter plasmid and 0.1 g of CMV G4 derivatives. C33A cells were cotransfected as in B. Transfection efficiency was routinely monitored by cotransfected ␤-galactosidase activity, and activation folds were calculated relative to luciferase activity in the presence of blank vector, which was set as 100%. Data presented are an average of three independent experiments, and standard deviations are indicated with error bars. mechanism involving the recruitment of HP1 and subsequent heterochromatinization of the promoter. We initially examined physical interactions between KSHV LANA and HP1 by using in vivo and in vitro binding assays. 293T cells were cotransfected with FLAG-tagged LANA and either a green fluorescent protein (GFP)-tagged human HP1␣ (hHP1␣) expression vector or an appropriate blank vector. After 36 h of transfection, cell extracts were prepared and immunoprecipitated with an anti-FLAG antibody. GFP-tagged hHP1␣, but not GFP alone, precipitated only in the presence of FLAG-tagged LANA, confirming physical interactions between the two proteins in mammalian cells (Fig. 2A). To map the LANA-binding region within hHP1␣, mammalian expression vectors encoding wildtype hHP1␣ or deletion mutants fused to GST were constructed (Fig. 2B, top) and cotransfected with the FLAG-tagged LANA expression vector into 293T cells. Proteins precipitated from FIG. 2. In vivo and in vitro association of KSHV LANA with HP1. A, in vivo coimmunoprecipitation of FLAG-tagged LANA and GFP-tagged hHP1␣. FLAG-tagged LANA and GFP or GFP-tagged hHP1␣ expression vectors were cotransfected into 293T cells. Following cell lysis, proteins immunoprecipitated (IP) with anti-FLAG antibody were immunoblotted with rabbit polyclonal anti-GFP serum (top). Total lysates from transfected cells were also immunoblotted with rabbit polyclonal anti-GFP serum (middle) and anti-FLAG antibody (bottom). B, in vivo GST pull-down assay between FLAG-tagged LANA and GST-hHP1␣ derivatives. FLAG-tagged LANA expression vector was transfected into 293T cells, along with expression vectors of GST or GST-hHP1␣ derivatives (top). After cell lysis, proteins pulled down by glutathione-Sepharose beads were immunoblotted with anti-FLAG antibody (top) and anti-GST antibody (middle). Total lysates from transfected cells were also immunoblotted with anti-FLAG antibody (bottom). CD, chromo domain; CSD, chromo shadow domain. C, in vivo GST pull-down assay between GST-LANA and FLAG-tagged HP1 isotypes. Expression vectors of GST or GST-LANA and indicated HP1 isotypes were cotransfected into 293T cells. Proteins precipitated with glutathione-Sepharose beads were immunoblotted with anti-FLAG (top) and anti-GST antibodies (bottom). Total lysates from transfected cells were additionally immunoblotted with anti-FLAG antibody (middle). D, Coomassie Blue staining of indicated GST fusion proteins at the top, which were bacterially expressed, purified, and used for in vitro binding. Protein size markers are shown on the left of the Coomassie stain panel, and GST fusion proteins with appropriate sizes are indicated by asterisks. E, binding assay between in vitro translated LANA and GST-HP1 isotypes. In vitro translated LANA in the presence of [ 35 S]methionine was incubated with glutathione-Sepharose 4B precoated with indicated GST fusion proteins. Where specified, 1 mM ATP was included in the binding buffer. Bound proteins were subjected to autoradiography. Input, 10% in vitro translated protein.
cell extracts with glutathione-Sepharose 4B were immunoblotted with anti-FLAG or anti-GST antibodies (Fig. 2B, bottom). As depicted in Fig. 2B, FLAG-tagged LANA was pulled down by GST-hHP1␣ and GST-hHP1␣ CD, but not GST alone or GST-hHP1␣ CSD, suggesting that LANA binds to the N terminus of hHP1␣ containing a chromo domain.
To date, three isoforms of HP1 have been identified in mammals (8 -11). Despite similar modular structures, the proteins display different subnuclear localization and phosphorylation patterns (12). Recent reports (66) indicate that HP1␣ and HP1␥ (but not HP1␤) specifically interact with human TAF II 130, suggesting that each HP1 isotype participates in distinct chromosomal and transcriptional events. We determined the isotype specificity of interactions between KSHV LANA and HP1. The GST or GST-fused LANA expression vector was cotransfected with the indicated FLAG-tagged HP1 expression vector in 293T cells, and a GST pull-down assay was similarly performed using extracts of transfected cells. FLAG-tagged HP1␣, but not mouse HP1␥, was efficiently precipitated with GST-LANA (Fig. 2C). FLAG-tagged mouse HP1␤ also weakly bound GST-LANA. To further confirm this isotype-specific interaction between KSHV LANA and HP1, isoforms of GST-fused HP1 were bacterially purified, and equivalent amounts were used in in vitro GST pull-down assays (Fig. 2D). In vitro translated LANA in the presence of [ 35 S]methionine was included in the binding reaction with indicated GST fusion proteins. After precipitation using glutathione-Sepharose 4B, bound proteins were subjected to SDS-PAGE and autoradiography. Our data reveal that only HP1␣ interacted with LANA. Furthermore, this interaction was enhanced in the presence of 1 mM ATP in the binding buffer (Fig. 2E).
To determine the HP1␣-interacting domain of LANA, a series of deletion mutants was constructed, translated in vitro, and employed in similar in vitro binding assays. The HP1␣interacting domain of LANA was mapped to the C-terminal region (Fig. 3). By using small truncation and internal deletion mutants of the C terminus of LANA, we demonstrated that the region encompassing amino acids 1047-1062 of LANA is responsible for interactions with HP1␣. Interestingly, the region required for HP1␣ binding coincided with the transcriptional repression domain of the LANA C terminus, with the exception that amino acids 1063-1162 did not interact with HP1␣ but displayed moderate transcriptional repressor activity.
The results collectively confirm that KSHV LANA interacts with HP1 proteins in an isotype-specific manner. Moreover, the region required for transcriptional repression of the LANA C terminus is necessary for binding to HP1␣, supporting the hypothesis that transcriptional repression activity of the C terminus of LANA may be at least partly mediated by recruiting HP1 to the promoter.
Subnuclear Localization of KSHV LANA and HP1-KSHV LANA displays characteristic speckled nuclear localization in latently infected cells, and this punctate distribution is mediated by the C terminus (37,59,65,67). HP1 also displays a distinct cell type-and isotype-specific subnuclear localization pattern (12,61,68). We speculate that the subnuclear distribution of LANA, and possibly association with heterochromatin, is mediated by interactions with HP1. Immunofluorescence assays were employed to determine whether the two proteins colocalize in cotransfected cells. LANA and FLAG-tagged expression vectors encoding the indicated HP1 isotypes were cotransfected into 293T cells, and subnuclear localization was detected using rabbit polyclonal anti-LANA serum (green) and mouse anti-FLAG M2 monoclonal antibody (red), respectively. As shown in Fig. 4, LANA and hHP1␣ clearly colocalized into relatively large subnuclear dots, although their punctate pat- terns did not accurately coincide. In contrast, we observed marginal or no significant colocalization of LANA with mHP1␤ and mHP1␥. These results are consistent with the in vivo and in vitro binding data and further support the hypothesis that the punctate subnuclear distribution of LANA may be partly mediated by associations with HP1.
Subcellular Fractionation of C-terminal Deletion Mutants of LANA Fused to GAL4 DBD-Because confocal microscopy data revealed that LANA and HP1 colocalize into specific subnuclear structures with characteristic speckled spots, we speculated that the transcriptional repression domain of the LANA C terminus correlates not only with the HP1␣-binding region but also with the region responsible for targeting to subnuclear structures. To verify this theory, we biochemically fractionated total proteins from transfected 293T cells expressing GAL4 DBD fusion proteins, which were used in the transient reporter assay to identify the transcriptional repression domain within the C-terminal LANA (Fig. 1D). A schematic procedure for subcellular fractionation is depicted in Fig. 5A. As shown in Fig. 5B, GAL4 DBD alone displayed little 0.1% Triton X-100extractable fraction, but the majority was equivalently observed in both DNase I-extractable and DNase I-resistant fractions. Fractionation patterns of a series of LANA C-terminal deletion mutants fused to GAL4 DBD revealed that the targeting signal for association with DNase I-resistant structures resides in amino acids 1026 -1062 of LANA, which includes the HP1␣-binding region. Similar results were obtained using 0.5% Triton X-100 instead of 0.1% (data not shown). This finding supports the possibility that targeting the LANA C terminus to subnuclear structures is mediated by associations with HP1 protein.
Transient Replication Assay Using KSHV oriP-containing Plasmid and hHP1␣-LANA Hybrid Proteins-Previously, we characterized the requirement of cis-and trans-elements of KSHV latent replication using a transient replication assay with a methylation-sensitive restriction enzyme (45). To investigate the functional role of interactions between LANA and HP1 in KSHV latent replication, we performed similar transient replication assays. Transient overexpression of hHP1␣ had no significant effect on the replication of p4TR-luc containing 4 tandem KSHV terminal repeats (TRs) by LANA in conditions where the overexpression of other trans-elements affected transient replication by LANA (data not shown). Because all internal deletion mutants of the C terminus of LANA failed to bind KSHV TR in an electrophoretic mobility shift assay, we did not include HP1-binding defective mutants of LANA in the transient replication assay, as in the case of origin recognition complexes binding defective mutants to check their ability in KSHV latent replication (45). Therefore, we constructed two-hybrid proteins in which the N or C terminus of hHP1␣ containing a chromo or chromo shadow domain, respectively, was fused to the C terminus of LANA that binds KSHV TRs and ORCs/HP1 to mediate the replication of a TR-containing plasmid (Fig. 6A), and we examined their contribution to KSHV latent replication. Appropriate sizes and comparable levels of FLAG-tagged CD-LANA C and CSD-LANA C were verified by Western blotting of total cell extracts from transfected 293T cells with an anti-FLAG M2 antibody (data not shown). Subcellular localization was determined by immunofluorescence assays using the same antibody. LANA-(⌬91-949), which mediates KSHV TR replication comparable with full-length LANA in transfected 293T cells (45), localized to the nucleus with a speckled pattern (Fig. 6B). Similar to full-length HP1, CD-LANA C also displayed more punctate subnuclear localization than LANA C alone. However, CSD-LANA C was more heterogeneous in the population of transfected cells and localized to both the nucleus and cytoplasm.
Previously, we reported that p4TR-luc containing KSHV TR, but not pGL2-basic, replicates only in the presence of LANA in transiently transfected 293T and BJAB cells (45). Similar results were obtained with the human cervical cancer cell line, C33A (data not shown). To determine whether chromo domain or chromo shadow domain of hHP1␣ fused to the C terminus of LANA supports the replication of TR-containing plasmids, expression vectors of these hybrid proteins along with p4TR-luc and pGL2-basic were cotransfected into C33A cells, and a transient replication assay was performed (Fig. 6C). Unexpectedly, TR-containing plasmids weakly replicated in the presence of LANA-(⌬91-949), compared with full-length LANA, in contrast to data obtained from 293T cells. However, LANA C alone did not support p4TR-luc replication, similar to the behavior observed in 293T cells. Interestingly, CD-LANA C showed the replication activity of TR-containing plasmids (albeit very weakly), compared with CSD-LANA C and LANA C alone. The

FIG. 4. Colocalization of LANA and HP1 isotypes in 293T cells.
Cells grown on coverslips were cotransfected with 1 g of pcDNA3 LANA and 1 g of ME18S-F HP1 derivatives expressing the indicated FLAG-tagged HP1 isotypes (left column). At ϳ24 h post-transfection, cells were fixed and immunostained. LANA was detected using a fluorescein isothiocyanate-conjugated secondary antibody against rabbit polyclonal anti-LANA serum (green), and FLAG-tagged HP1 isotypes were identified using a rhodamine-conjugated secondary antibody against anti-FLAG monoclonal antibody (red). Two representative confocal microscopic images per HP1 isotype are shown.
internal control, pGL2-basic, was employed to confirm complete digestion of unreplicated DNA by DpnI. The expression level of the trans-acting element was largely excessive, which excluded the possibility that distinct replication activities of each mutant were due to different expression levels of them. The results imply that the N terminus of hHP1␣ containing a chromo domain functionally replaces the N terminus of LANA containing a chromosome-binding domain, at least in part, and support the possibility that associations between LANA and HP1 contribute to the replication activity of LANA.

DISCUSSION
The transcriptional repression activity of LANA, including LANA derivatives fused to GAL4 DBD and reporters with tandem repeats of GAL4 DBD-binding sites, was previously identified from transient reporter assays (50,59). LANA represses the transcriptional activity of KSHV TR by direct binding (43)(44)(45). Transcriptional repression activity of C-terminal LANA fused to GAL4 DBD was controversially observed, although the C terminus alone was necessary and sufficient for transcriptional repression of KSHV TR (43)(44)(45). We further characterized the transcriptional activity of LANA derivatives in the GAL4 system. A transient reporter assay using two cell lines and four different reporters revealed cell type-and promoter-specific repressor activity of LANA derivatives fused to GAL4 DBD (Fig. 1). In view of earlier reports that the N terminus of LANA interacts with and represses the mSin3 corepressor complex (50), we were interested in determining the mechanism of transcriptional repression by C-terminal LANA.
Association of LANA with heterochromatin (65) and localization of the C-terminal region to subnuclear structures (59) suggested a possible link between LANA and HP1, a representative heterochromatin-associated protein. In vivo and in vitro binding assays revealed isotype-specific interactions between the two proteins ( Figs. 2 and 3). Immunofluorescence data using cotransfected 293T cells further supported the binding assay results (Fig. 4). A targeting signal to a DNase I-resistant structure was also biochemically identified within C-terminal LANA (Fig. 5). Data from a series of LANA deletion mutants, in addition to our recent finding that C-terminal LANA interacts with ORCs (45), are summarized in Table I. The HP1-and ORC-binding region of LANA was determined from experiments with LANA- (⌬1002-1062). The mutant did not bind HP1 or associate with DNase I-resistant structures but interacted with ORCs and possessed moderate transcriptional repression activity, supporting the hypothesis that transcriptional repression and targeting to subnuclear structures is a result of associations with HP1 protein. Because ORCs showed the transcriptional repression activity of reporter when tethered to a heterologous promoter via GAL4 DBD (data not shown), the residual transcriptional repression activity of LANA-(⌬1002-1062) may be explained by interactions with ORCs.
In view of the finding that LANA binds KSHV TR and exerts transcriptional repression and replication activity, the association between LANA and HP1 presents an attractive model. LANA may repress transcription from KSHV TR, at least in part, by recruiting HP1 proteins so that the gene-silencing mechanism works in heterochromatic chromosomes. To investigate this possibility, we examined the effect of overexpressed HP1␣ on the transcriptional activity of p4TR-luc in the presence or absence of LANA. In our transient reporter assay, overexpressed HP1␣ did not have a significant effect on transcriptional repression of p4TR-luc by LANA, compared with basal transcriptional activity of KSHV TR in the absence of LANA (data not shown). A transient reporter assay using the GAL4 system revealed similar data. It is possible that interactions with other factors, such as ORCs, play a dominant role in the transcriptional repression activity of LANA. Alternatively, endogenous HP1 may be abundant in transfected cells relative to exogenously expressed protein, or only a small fraction of HP1 may participate in transcriptional repression by LANA. A transient reporter assay involving this type of overexpression system did not give conclusive results on the effects of HP1 on transcriptional repression by LANA at the physiological level, as discussed by Vassallo and Tanese (66).
Although we could not experimentally prove the heterochromatinization of KSHV TR by LANA via possible interactions with HP1, this model benefits viral infection in several ways. KSHV contains 30 -40 copies of the TR sequence within its genome. This tandem repeat sequence is more prone to gradual loss during DNA replication or recombination events. Therefore, KSHV must have a protection mechanism to prevent the loss of TR, which acts as a latent origin of replication required for maintaining the genome. Heterochromatinization of TR FIG. 6. Transient replication assay of KSHV TR-containing plasmids, p4TR-luc, using hHP1␣-LANA C hybrid constructs in C33A cells. A, schematic diagram of LANA deletion mutants and hHP1␣-LANA C hybrids. CD, chromo domain; CSD, chromo shadow domain. B, subcellular localization of LANA deletion mutants and hHP1␣-LANA C hybrids in 293T cells. Cells grown on coverslips were transfected with 1 g of FLAG-CMV2 derivatives expressing FLAG-tagged LANA derivatives and hHP1␣-LANA C hybrids. At ϳ24 h post-transfection, cells were fixed and immunostained. FLAG-tagged proteins were detected using an anti-FLAG monoclonal antibody and a rhodamine-conjugated secondary antibody. C, transient replication assay in C33A cells. Cells in 100-mm dishes were cotransfected with 2 g of p4TR-luc, 2 g of pGL2-basic, and 8 g of the indicated FLAG-CMV2 derivatives. Cells were split at ϳ36 h post-transfection and harvested at ϳ96 h. Hirt-extracted DNA from transfected C33A cells was digested with Alw44I/DpnI (right), and 20% samples with Alw44I alone as input (left). Digested DNA was separated by 0.8% agarose gel electrophoresis and analyzed by Southern blot hybridization with a probe specific for the luciferase gene. may be one solution. In addition, targeting of the viral genome to subnuclear structures, coupling of viral replication to that of heterochromatin by associations between LANA and HP1, or the centromere function of HP1 (69) may partly contribute to viral genome replication, which was preliminarily verified by the transient replication assay using HP1-LANA C hybrids in this report (Fig. 6). With respect to transcriptional regulation, heterochromatinzation of TR may affect the activity of other regions within the viral genome by intramolecular contacts. This type of trans effect (70) may globally restrict transcriptionally active viral promoters during the latent infection cycle or in the absence of appropriate signals. More detailed studies are required for validating these models and possibilities.